The Agent Skills Directory

[SAFE]: The skill 'evo-memory' manages persistent research knowledge by reading experiment logs and idea summaries to update markdown-based memory stores within the local /memory/ directory. This activity is consistent with its stated purpose of meta-learning across research cycles.
[SAFE]: The tool configuration (allowed-tools) is restricted to basic file operations and internal reasoning. No tools for network communication, subprocess execution, or administrative privilege escalation are requested, minimizing the attack surface.
[SAFE]: The instructions and prompts for the evolution mechanisms (IDE, IVE, ESE) are focused on technical analysis and technical data extraction. No evidence of prompt injection or attempts to bypass agent safety filters was found in any of the skill's reference or asset files.
[SAFE]: The skill operates entirely within the workspace and does not establish external persistence or download third-party dependencies. All logic is implemented via natural language instructions and structured templates, with no executable scripts or binary code included.

evo-memory